Comparison of two data sets at different levels/options of the tuna atlas

The purpose of this markdown is to describe the differences of data from the World Tuna Atlas following the different filters. It describes the differences between the initial data and the final data, and is intended for users of the final data who might be tempted to use it without taking into account the various filtering specificities.

Attention ! All the differences inferior to 0 corresponds to gain in captures.

1 The data compared

The analyzed data are :

• acces rawdata_modyfing_georeferenced_errors for the initial data

• acces treatment_after_binding for the final data

There are no filter on this data

The filter used on this data are:

• on species:
• on gears:
• on rfmos:
• on fleets:
• on initial dates:
• on final dates:
• on geographical categories:
• on catch types:

There are no differences between the options to create the two datasets.They are created by the same workflow.

The options used to create the final data frame are:

2 Main differences

The number of lines goes from 4.679497 millions in rawdata_modyfing_georeferenced_errors to 4.703429 millions in treatment_after_binding, which correspond to a difference of -0.51142%.

The initial dataset has a total of 103363606 in tons and of 1148126339 in number of fish.

The final dataset has a total of 103386195 in tons and of 1148126339 in number of fish.

The differences is -22589 in tons (-0.02%). The differences is 0 in number of fish (0).

The stratas differences between the first one and the second one are :

(only first 10 per Dimension showed), representing 0 % of the total number of stratas.

3 Introduction to the two datasets

We first present the main characteristics of each dimension for each dataset.

3.1 Comparison of the two catch evolutions

For each dataset, we compare the catch evolution and the cumulative catch evolution both in tons and number of fish.

3.2 Comparison of the evolution of the cumulative catches

3.3 Comparison of the two datasets

We check the distribution of the value of each dimension in tons and number of fish for each dataset.

3.4 Spatial coverage

We represent the spatial coverage, faceted by the geographical category. The geographical category depends on the area of the geographic polygon. In this case there are categories which are . The category is made on the area, also different shapes could be included in the same geographical category (example 5x10, 10x5 and 25x2).

3.4.1 Spatial coverage in tons

3.4.2 Spatial coverage in number of fish

4 Differences

This section detail the different differences that observed between the dataframe rawdata_modyfing_georeferenced_errors and treatment_after_binding.

4.1 The differences for each dimension

We will look for each dimension the 6 most important differences

4.2 Differences in geographical data

Here is represented for each area the polygons keeping all the initial information, the one losing a part and the one losing all the information. The polygongs with red bord lost all the information, the one with blue borders did not loose information.

The coverage difference is -1 km square.

4.3 Differences in temporal data

Here is represented the differences in percent for each year.